This is part of the Michael H. Moskow Honorary Paper Series, which also includes:
Early in 1973, Michael Moskow asked me to be one of his deputies in the Office of Policy Development and Research at the Department of Housing and Urban Development (HUD). Given his reputation, I eagerly accepted. Jim Lynn was the newly appointed HUD secretary and although I did not know him, I had watched him perform in White House meetings when I was at the Council of Economic Advisers. I was very impressed.
Not long after my arrival I found myself responsible for the Experimental Housing Allowance Program, a huge social experiment involving twelve cities and 20,000 recipients. It was the most daunting management challenge of my life. The late Jerry Fitts, an extremely talented civil servant who had learned his management skills at NASA, was an invaluable partner.
Two cities — Phoenix and Pittsburgh — were chosen for an experiment, known as the demand experiment, that tried out different formulas to subsidize vouchers for a sample of low-income households wanting to rent existing housing. Green Bay, Wisconsin, and South Bend, Indiana, were chosen for an experiment called the supply experiment that asked whether there would be a sufficient supply of housing available if demand was increased by offering such subsidies. There, all those eligible for vouchers were provided subsidies, so that we could investigate the effects of a city-wide increase in housing demand. In eight other cities we tried different approaches in administering the vouchers. In some, states administered the program; in others, counties were used; and in still others we relied on housing authorities.
In today’s era of extremely tight budgets for discretionary programs, it is impossible to imagine that such a large, expensive, social experiment would be funded by appropriation committees. But they ploughed ahead, even though the program was highly controversial.
The design of the experiment was finalized long before I arrived at HUD. I was shocked to find that the size of the subsidy was primarily based on a recipient’s income. The maximum subsidy depended on rental prices in a locality, but for recipients receiving less than the maximum, the size of the subsidy was not related to the cost of housing. It seemed more accurate to call it an income supplement program instead of a housing program. However, it had some housing elements. There were housing codes to weed out undesirable units and recipients did receive counselling to explain their rights as renters and the responsibilities of landlords.
The program was controversial for two main reasons. Housing subsidies had historically relied mainly on newly constructed housing. Lobbyists for the housing construction industry rightly feared losing business and they vigorously argued that the provision of vouchers to renters of existing housing would create housing shortages and drive up rents. The other controversy had a racial element. People in better neighborhoods, especially in the suburbs, worried that vouchers would finance an invasion of minorities that would depress the value of their homes.
The latter fear, bolstered by people who just did not like welfare programs of any type, made it difficult to find cities that would accept the supply experiment. Green Bay had willingly accepted it before I got to HUD, but I was left with the task of trying to persuade the good citizens of Springfield, Ohio, to accept HUD’s money. We assured them that the vouchers would strengthen the local economy. They would have none of it. I appeared at a town meeting where the opposition was so boisterous that I feared for my physical safety. We tried a number of other cities and finally South Bend, Indiana, accepted.
The findings of the experiments were fascinating. Most interesting, the typical voucher recipient did not move very far, often not more than a few hundred yards. Moves to the distant suburbs were extremely rare and sometimes people stayed in the same unit. The counselling taught them that they could demand repairs from their landlords like fixing broken windows and holes in screens.
The Green Bay and South Bend experiments proved definitively that there was a sufficient supply of housing in local markets to satisfy the extra demand created by vouchers. In other words, the vouchers did not cause significant rent increases. Consequently, there was no need for new construction which, in any case, was more expensive per subsidized unit. Housing construction advocates have never accepted these findings. When I appeared a few years ago at a meeting introducing an excellent book on housing policy by John Weicher, advocates of new construction were still arguing with proponents of vouchers.
Those of us involved in the housing experiment and in an accompanying exhaustive and exhausting housing policy study hoped that the experiment could serve as a model for a nationwide, entitlement voucher program. But that was an impossible dream. It would have been far too expensive. Instead the results of the experiment were extremely helpful in designing a new program with a very limited budget that financed vouchers that helped low-income households rent existing housing. The program, called Section 8, also had a new construction component, showing that the housing construction industry still had plenty of political power.
Because of limited budgets only a small portion of the eligible population can be served. Today, only one quarter of the eligible population gets vouchers and there are long queues. Sometimes no new subsidies are available in certain cities for years.
Public finance economists of all ideologies believe that people that are equal should be treated equally by the tax-transfer system, although they may differ on how to measure equality. Most refer to income as the appropriate measure, while some prefer consumption, and a few may favor wealth. While at HUD, several of us tried to persuade congressional staff that with a limited budget it might be better to provide a shallower subsidy that could serve more people regardless of the measure used to define equality. I quickly learned that politicians prefer to provide a big subsidy to a limited number of households, so that they are well-remembered by the favored few.
I have emphasized the Section 8 program because it was designed while Mike and I were at HUD. That program mainly replaced two disastrous homeownership and rental programs know as Sections 235 and 236. They were designed so poorly that they resulted in a huge number of mortgage defaults and foreclosures. Whole neighborhoods were destroyed. Detroit suffered disproportionately.
HUD still has an astounding number of programs, each designed a little differently and intended to serve a slightly different purpose. Among others, there are programs that encourage home ownership, programs that can involve religious groups, and a generous tax credit program that encourages the construction of low-income housing. This last tax credit survived the recent tax reform.
The essence of a voucher is that it expands the choices available to recipients. They do not have to take it to a particular building in a particular place to enjoy the subsidy. The choice is limited by a housing code applied nationally to the whole program, but not by much else. Vouchers for primary and secondary schools are similar and food stamps, now known as SNAP, also have only a few restrictions. If freedom of choice is an important goal, one might ask, “Why not go all the way and just give poor people cash?” However, that question would open a much larger discussion.
HUD has been recently experimenting with a restriction on vouchers that would require recipients to use them to move to better neighborhoods, measured mainly by how many in the neighborhood have gone to college. The results are being studied by Raj Chetty, Nathaniel Hendren, and Lawrence Katz in conjunction with MDRC, a nonprofit company that carries out randomized, controlled experiments to evaluate social programs. (Full disclosure: I am on the board of MDRC together with Larry Katz.) The results, which may still be preliminary, suggest that the experiment, which is an elaboration of the housing allowance experiment carried out while Mike and I were at HUD, is remarkably successful. The income of adult recipients is up 31 percent compared to those in the traditional program; single parenthood is down 30 percent; and college attendance is up 30 percent. It suggests that behavior is very much affected by those living close to you.